In the scientific world, theories are not just whatever explanation someone dreams up. An idea must be supported by data and observations or it is thrown out before ever becoming a theory. There is a clear difference between the scientific definition of a theory and the common use of the word. A scientific theory has been verified against the data and is generally accepted as an accurate explanation of the observations. Better descriptions that I can do are here and here.
But a theory is not fixed and final. Theories are often modified as new discoveries are made. That is one of the key ways that science continues to move forward. A theory is updated until it can not be changed to fit the data, then a new theory is started. And even then the previous theory may not be invalidated, it is just modified to only apply to certain cases. There have been a couple of recent high profile cases where some claimed we should abandon a theory based on one piece of evidence that does not fit. But that is not at all how science works. A theory is only thrown away when another comes along which explains more, and does it more accurately. They are not discarded without a scientifically proven replacement that does the job better, even if we know they are not yet perfect on their own.
Often, theories remain even after they have been disproved. Though difficult to imagine, science is not very concerned with the absolute truth. Instead it looks for the ideas and methods that most accurately predict what happens. For example, say scientists are trying to determine what is that the center of a black hole. No light can escape so we can not observe it directly. If scientists were only concerned with the truth, they would abandon research, because finding the true answer will likely be impossible. But what scientists are really looking for is some explanation of how the black hole acts that agrees with all of the data, regardless of whether their idea is actually what they would find in the center of the black hole.
The best theories are ones that fit what we observe in the universe, not some kind of universal truth. It is the reason the Bohr model is still taught in high school science, even though it was proved obsolete by quantum mechanics. Then there is my favorite example: gravity. Guess what, Newton's theory of gravity is wrong. It does not work at very small or large distances or high speeds. Einstein's general relativity replaced it, but even that did not make it irrelevant as a teaching tool or an easy way to make most calculations for the conditions we deal with.
The point here is that science looks for the methods that work best for predictions, however well that may be. Theories are not thrown away with no replacement if they are found to be imperfect. We will continue to teach those that are most useful, and when science develops a theory, know that is has been tested against the data. But even then, that theory may be universally accurate, or it may only work under a small set of conditions. So the phrase "Just a theory" is not just misused on scientific ideas, it really does not mean anything at all. The only thing to throw out is any argument that begins with that phrase.
